Central Bank of Ireland ETF Discussion Paper Response

Central Bank of Ireland ETF Discussion Paper Response August 2017

Introduction Thank you for the elaborate, well-researched Discussion Paper on Exchange Traded Funds and for giving us the opportunity to respond to it. As a global, leading liquidity provider specializing in ETFs we are an Authorized Participant (AP) and Liquidity Provider (LP) for a large number of ETFs in Europe and are pleased to contribute our views in respect of the ETF product and ecosystem. Background Flow Traders was founded in Amsterdam (the Netherlands) in 2004 and our trading desks in Europe, the Americas and Asia provide liquidity across all major exchanges, globally, 24 hours a day. We use our proprietary technology platform to quote bid and ask prices in thousands of ETP listings, as well as similar financial products. We also provide liquidity to institutional counterparties off-exchange across all regions. As a principal trading firm, we trade for our own account and own risk only. We do not have clients, nor do we provide any investment services or ancillary services to others. Our strategies are designed to use information that is publicly available and we use fairly simple, non-controversial and transparent order types. We are a strong supporter of fair, transparent and orderly markets. Flow Traders is regulated in the Netherlands (DNB and AFM) and the US (SEC and FINRA) and has been listed on Euronext Amsterdam since 2015. Overview The most important feature of the ETF product is that it is highly transparent in terms of composition of the underlying assets and costs, while the most important feature of its ecosystem is that it is open, competitive and redundant. These built-in elements promote that ETFs can be traded cheaply and at a high liquidity and that the ecosystem is resilient. If the composition of an ETF is known to the market, or at least all LPs, price discovery is relatively straightforward. Market participants can verify if pricing is fair and any deviation from NAV will be arbitraged away provided that market participants, LPs in particular, can act in a competitive environment and have sufficient information. This holds even in stressed markets although the spread will widen in order to account for the LPs risk premium in such circumstances. Having multiple APs and LPs (and other supporting roles in the ETF s infrastructure) is essential in such constellation: dispersing roles over multiple parties (rather than an infrastructure dominated by the issuer and its group) reduces single points of failure and at the same time promotes competitive pricing. This benefits investors and market quality at large. Conversely, would an ETF have a single AP/LP, particularly where such AP/LP is affiliated with the issuer (incumbent), these essential features diminish exposing markets/investors to potential discretionary behavior or increased risk of failure. Therefore, AP- and LP-ships should be open to all qualifying market participants, independent from the issuer, in order to foster competition, better pricing (closer to NAV) and resilience. In addition, any appointed LP should commit to substantial obligations in terms of time-in-market (often 95%), maximum spread and minimum volume in respect of their quoting in order to ensure market quality. The minimum requirements set out in the market making obligations pursuant to MiFID II will not suffice. Flow Traders ETF Discussion Paper Response 2

ETF infrastructure transparency and resilience (Questions A, G and H) A key theme in the Discussion Paper is the dependence upon the AP infrastructure in stressed markets, particularly where the number of APs is limited or reduced to one. We believe that appointing multiple APs and LPs, trading in competition, improves market quality and reduces the ETF infrastructure s risk and costs. AP and LP roles should not be held by parties affiliated with the issuer only, but ideally be open to any qualifying party. In addition, transparency is an essential feature of the ETF product and ecosystem. Investors should know what is contained in their ETFs (although some products might deviate from that premise if managed well, e.g. actively-managed ETFs) and who are involved in its infrastructure. Most exchanges disclose which APs and LPs are appointed for the relevant ETFs, which we see as good practice. Transparency in respect of aggregate fees paid to the relevant roles in the ETF s infrastructure is highly relevant information. Investors should know what they pay for. However, overly detailed disclosure of the fee structure is of reduced relevance to the resilience of the ecosystem, and fee arrangements are commercially sensitive. A balance needs to be struck: As long as terms between the issuer and its supporting parties are competitive, at arm s length, with multiple nonincumbent parties (i.e. not a related entity or affiliate as AP/LP), aggregate/high-level disclosure of arrangements and payments to supporting parties should suffice. More detailed data relating to individual parties in the ETF s infrastructure becomes commercially sensitive to both the issuer and APs/LPs. Seeding arrangements or LP arrangements for newly launched or less liquid ETFs can be commercially sensitive to the issuer. Additionally, APs and LPs typically operate on similar terms but providing liquidity is a highly competitive business with very low margins, so overly detailed disclosure is also sensitive to APs/LPs. Finally, information beyond aggregate fee disclosure does not provide investors with any additional meaningful information or promote resilience. Most relevant to disclose are payments to, arrangements with, and dependencies upon, parties affiliated with the issuer. Additionally, there should be full disclosure where a single point of failure exists in the structure (e.g. non-redundancy in terms of the AP and LP function, or swap providers). It is very important for the market to know where such arrangements expose investors to inflated costs, issuer affiliate discretion, potential conflicting interests, or risks inherent to failure of any such party. An issuer appointing affiliated parties in its ETF s infrastructure (aside competing APs/LPs) is not necessarily problematic but any concentration of roles, or any potential conflict of interest, should be fully disclosed in order to allow monitoring. Functioning of the ETF s infrastructure under distress (Questions C, D and N) Investors must be made aware that ETFs could temporarily become closed-ended and that pricing will typically be off- NAV as a result of extreme market circumstances. Even when an ETF would temporarily be unable to accept creations or redemptions/become closed-ended, market participants will still be able to trade the ETF throughout the day albeit at a wider spread, still allowing risk transfer among investors. As shown during extreme markets over the last decade, LPs will likely remain in the market as it may provide lucrative trading opportunities. At such moments, they do seek compensation for assuming higher risks. This is reflected in the spread Flow Traders ETF Discussion Paper Response 3

they quote. This does not necessarily imply that the price (or spread) is wrong or unfair, given sufficient competition: as the best quote wins (at the expense of the others), even when spreads are wide, markets still provide competitive incentives particularly when LPs can manage their risks and positions if they have access to the creation/redemption process. In that light, appointing multiple LPs should be promoted and APs/LPs should not be limited to parties incumbent/affiliated with the issuer. This should deter discretionary behavior or limited competition, which could hamper efficient price discovery and liquidity provision. It would be practically challenging to open up the creation/redemption process to non-aps. The open ETF infrastructure implies that multiple professional parties, in competition, improve the ecosystem s resilience even under stressed circumstances. Issuers are typically not set up to accept creations and redemptions from a wide range of investors. Concentrating that function with APs reduces the ETF infrastructure s costs. However, in case of a collapsing ETF or if its supporting infrastructure would significantly be impaired (despite the fact that the setup is designed to be redundant), redemption for a wider range of investors could be an option of last resort. Transparency regarding ETF composition (Questions B, J and K) A key theme addressed in the Discussion Paper is transparency in respect of the composition of ETFs. Such transparency allows investors to assess their exposure, but also serves to promote effective and efficient price discovery, allowing for arbitrage where ETF prices are deviating from their NAV. As a matter of principle, ETF composition should be disclosed as transparently, fully and timely as possible. Less (or delayed) disclosure typically results in less liquidity and/or higher spreads. Full and immediate disclosure may pose challenges for actively-managed ETFs. In protecting their intellectual property, we understand that issuers may be reluctant to disclose the holdings of their actively-managed ETFs. We believe that active ETFs may serve a valid purpose as a less expensive alternative for mutual funds, tradeable throughout the day. Nevertheless, we believe that such ETFs should retain the transparency and open infrastructure that so strongly underpins the ETF market, to the benefit of market quality and investor protection. Actively-managed ETFs should clearly disclose their sought exposure and strategy, as well as their actual exposure (bearing in mind commercial sensitivity). Inverse and highly leveraged products may pose high risks, particularly for retail investors, so transparency of composition and risks is essential. Arrangements with affiliates or others that could affect the investor in terms of costs, potential for discretion or conflict of interest and dependencies must also be fully disclosed and observed: investors must not be harmed by hidden arrangements, risks or costs, particularly in a product that is inherently more opaque than a passively-managed, fully transparent ETF. Delaying disclosure of composition limits investors in assessing the exposure they assume, at least temporarily. This is, however, not necessarily harmful to effective price discovery. As long as an actively-managed ETF (i) has multiple LPs appointed, and (ii) such LPs simultaneously receive full and up-to-date composition information (whether or not under NDA), any price differences will be arbitraged away competitively among such multiple LPs. Nevertheless, investors would still be uncertain as to the exposures they actually face at any given moment. Here, too, interests need to be balanced. Flow Traders ETF Discussion Paper Response 4

Relative liquidity of the ETF and its underlying assets (Questions L and M) The Discussion Paper addresses potential gaps between an ETF s market price, NAV and the value of underlying assets. An ETF s market price results from supply and demand. When an ETF is underpriced or overpriced compared to its underlying assets, market participants have an incentive to arbitrage such differences away. The spread may widen to reflect risks or limited liquidity so the ETF s market price does not always reflect its NAV. This does, however, assume that multiple APs/LPs are active, trading in competition. This mechanism is not necessarily depending on the liquidity or tradability of the underlying assets. LPs (and other market participants) are not limited to trading or hedging the underlying assets as they can hedge their ETF exposure through other instruments generating the offsetting exposure sought. Also when redeeming or creating, the bucket does not always reflect the full composition of an ETF. Particularly when underlying assets are illiquid or markets are closed, corresponding hedges can be found in instruments with similar exposure (often futures). If composition is known and competition is ensured, such proxy-hedging will typically push pricing close to NAV irrespective of relative liquidity or tradability of the underlying assets. Of course the spread will depend on the risks a liquidity provider assumes, which in turn is driven by market circumstances and competition. Some asset classes have lower inherent liquidity such as certain bonds and commodities. We see liquidity in such underlying asset classes improving, particularly where such instruments become more standardized and trading more competitively. The increased popularity of ETFs for such asset classes seems to have a positive impact. Furthermore, we hope to see more bond volume and derivatives moving from a restricted, closed, trading environment to more transparent and open, electronic, markets. We wish to stress that such increased competition is essential and should be promoted. This improves price discovery, transparency, market quality and resilience in markets for the underlying assets, positively influencing ETF market quality at the same time. We are less concerned that price discovery of underlyings is hampered when ETFs hold a significant or increasing proportion in their underlying assets. As long as free float remains sufficient, competitive markets ensure adequate price discovery, even where ETFs hold a large proportion of an underlying asset. Please note that roughly 6% of global assets under management is currently invested in ETFs. Sufficient alpha-seeking (active) strategies remain which drive supply and demand for individual underlying assets, regardless of a growing proportion of ETF investments merely following price levels set in underlying assets. Comparison with the US market structure and regulation beyond UCITS and MiFID II (Questions O and P) Market structure Although the US market is more mature and much larger, we do see similar market behavior and increasingly lower spreads in European markets. European markets appear to be operating competitively and quite resiliently, also in comparison with US markets. We remain strong proponents of open, transparent and competitive markets, preferably trading on lit venues such as Regulated Markets or MTFs. Any regulatory incentive to do otherwise, e.g. forcing market participants in club-like or dark market structures such as SIs are inconsistent with this policy aim and market quality, and should be closely monitored/discouraged. While we do see MiFID II as an important driver to promote trading on lit markets (e.g. Regulated Flow Traders ETF Discussion Paper Response 5

Markets and MTFs), which we strongly promote, MiFID II s provisions in respect of Systematic Internalisers may result in less fair, transparent and competitive markets than could have been the case otherwise. Settlement cycle and stock borrowing and lending An important feature of US markets is that the settlement cycle is extended for registered LPs that have quoting obligations on exchange. The creation and redemption cycle for some ETFs in Europe is T+3 while underlying assets have a mismatching settlement cycle of T+2. LPs trading from a short position, which they must have from time to time to in order to effectively provide liquidity, may be facing late delivery penalties and buy-ins, which increase costs and force the ETFs to be priced away from their NAVs. Providing LPs with an extension like in the US of e.g. T+5 or more will simplify operations and improve spreads by lowering costs. In addition, stock borrowing and lending markets could see some improvement in order to make them more effective and efficient (particularly compared to increased dynamism in markets for the ETFs themselves). It would be greatly beneficial for operational and pricing efficiency if ETFs would be accepted by CCPs as collateral for stock borrowing and lending transactions, which is currently not the case. Capital requirements We remain concerned that an upcoming capital requirements regime for investment firms may be inappropriate for liquidity providers in general, potentially increasing costs and affecting market quality. EBA s suggestions in this respect seem to rely on bank-like capital requirement concepts, which are not necessarily fit for purpose for liquidity providers. These proposals also seem to substantially overestimate required capital. Current suggestions do not seem to recognize that liquidity providers typically seek to hedge/set-off their portfolios to a large (if not full) extent, and elements in the framework appear to be incredibly complex and overestimating risks inherent to market makers. This is a reason for concern if not adequately addressed as it impacts markets as a whole. Liquidity providers are typically low-risk operations, setting off their positions to a large extent or in full. They do not hold or manage client assets or funds and clear and settle their trades through highly capitalized and regulated prime brokers. These prime brokers effectively assume all risks contained in a liquidity provider s portfolio (provided that the liquidity provider has no clients of its own and does not settle any trades outside its prime broker arrangements). The prime broker, in return, requires the liquidity provider to post collateral in cash (or equivalent), in order to cover for any potential risk such prime broker may face. In addition, when a liquidity provider fails, this has very limited impact on other market participants as the liquidity provision is a highly dispersed and competitive activity, and there are no deposit holders to protect: liquidity providers trade with their own funds, for their own risk. The most relevant prudential risk that needs to be covered is an orderly wind-down of a failing liquidity provider. All other risks are already effectively accounted for by posting collateral as required by the prime broker, extinguishing any market risk, liquidity risk, currency risk and all other relevant risks in a very conservative and prudent manner. Given liquidity providers low risk profile, lack of customer funds/deposits, non-systemic nature and very limited interconnectedness, a capital requirements regime should be proportionally lighter and less complex than capital requirements designed for banks. A regime that overestimates risks and requires inflated levels of capital will make it increasingly costly or practically cumbersome for liquidity providers to provide liquidity when markets need it most. **** Flow Traders ETF Discussion Paper Response 6